摘要
提出一种新型神经网络模型(HNM)·此模型是一种本质非线性模型,但可以应用线性控制理论的成果来设计稳定的控制器,并且模型中的连接权系数有各自的物理意义,可以通过经验来确定其初始值·同时通过分析对角矩阵的稳定性,给出了基于HN模型的控制器的设计方法及稳定性证明·仿真结果表明HN模型的有效性及控制器的优良性能·
?A new neural network model (Hyperbolic Network Model) was proposed. Although it is a nonlinear model in nature, the controller can be designed based on linear system theory. The factors have their own physical meaning and the initial factors can be determined by experience easily. The controller design method and stability analysis were proposed via the stability theory of diagonal matrix. An inverted pendulum system was used to illustrate the design procedure. The simulation results show that the HN model is suitable for modeling and control.
出处
《东北大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2002年第12期1131-1134,共4页
Journal of Northeastern University(Natural Science)
基金
国家自然科学基金资助项目(60274017)
教育部博士学科点专项科研基金资助项目(2001045023)